Numerous Arctic Ocean circulation and geochemical studies suggest that ice growth in polynyas over the Alaskan, Siberian, and Canadian continental shelves is a source of cold, saline water which contributes to the maintenance of the Arctic Ocean halocline. The purpose of this study is to estimate for the 1978--1987 winters the contributions of Arctic coastal polynyas to the cold halocline layer of the Arctic Ocean. The study uses a combination of satellite, oceanographic, and weather data to calculate the brine fluxes from the polynyas; then an oceanic box model is used to calculate their contributions to the cold halocline layer of the Arctic Ocean. This study complements and corrects a previous study of dense water production by coastal polynyas in the Barents, Kara, and Laptev Seas. Recurrent polynyas form on the Canadian and Alaskan coasts from Banks Island to the Bering Strait and on the Siberian coast from the Bering Strait to the New Siberian Islands.

In the Bering Sea, polynyas form in Norton Sound, south and west of St. Lawrence Island, and in the Gulf of Anadyr. Two regions that account for almost 50% of the total dense water production are the Siberian coastal polynyas in the adjacent regions of the Gulf of Anadyr and Anadyr Strait and the Alaskan coastal polynyas which occur along the coast from Cape Lisburne to Point Barrow. For all of the western Arctic coastal regions examined, the mean annual total brine flux is 0.5¿0.2 Sv. Combination of this flux with the contribution from the Barents, Kara, and Laptev Seas, which is recalculated from data in the earlier study, shows that over the entire Arctic, coastal polynyas generate about 0.7--1.2 Sv of dense water. This compares well with the theoretical estimates of 1--1.5 Sv. Because an unknown fraction of the Barents, Kara, and Laptev brine flux must go to the Eurasian Basin deep water, the coastal polynyas alone cannot renew the halocline layer.

Other potential brine generation mechanisms include overall freezing on the shelves and the response of the ice to infrequent violent storms. For example, during February 1982 an intense storm generated a large region of low ice concentration in the eastern Chukchi Sea over Barrow Canyon. The refreezing of the region was followed by the flow of a dense plume down Barrow Canyon. Although the ocean dynamical response to this refreezing needs to be established, the possible response of a Barrow Canyon flow to this refreezing event suggests that the overall refreezing in response to infrequent violent storms may be a potential source of the additional brine needed to maintain the Arctic Ocean halocline. ¿ American Geophysical Union 1994